Genetic and demographic signatures accompanying the evolution of the selfing syndrome in Daphne kiusiana, an evergreen shrub

Abstract

Background and aims: The evolution of mating systems from outcrossing to self-fertilization is a common transition in flowering plants. This shift is often associated with the 'selfing syndrome', which is characterized by less visible flowers with functional changes to control outcrossing. In most cases, the evolutionary history and demographic dynamics underlying the evolution of the selfing syndrome remain poorly understood.

Methods: Here, we characterize differences in the demographic genetic consequences and associated floral-specific traits between two distinct geographical groups of a wild shrub, Daphne kiusiana, endemic to East Asia; plants in the eastern region (southeastern Korea and Kyushu, Japan) exhibit smaller and fewer flowers compared to those of plants in the western region (southwestern Korea). Genetic analyses were conducted using nuclear microsatellites and chloroplast DNA (multiplexed phylogenetic marker sequencing) datasets.

Key results: A high selfing rate with significantly increased homozygosity characterized the eastern lineage, associated with lower levels of visibility and herkogamy in the floral traits. The two lineages harboured independent phylogeographical histories. In contrast to the western lineage, the eastern lineage showed a gradual reduction in the effective population size with no signs of a severe bottleneck despite its extreme range contraction during the last glacial period.

Conclusions: Our results suggest that the selfing-associated morphological changes in D. kiusiana are of relatively old origin (at least 100 000 years ago) and were driven by directional selection for efficient self-pollination. We provide evidence that the evolution of the selfing syndrome in D. kiusiana is not strongly associated with a severe population bottleneck.

Keywords: Daphne kiusiana; NGS technique; demographic dynamics; directional selection; evergreen shrub; inflorescence; selfing syndrome.

Fig. 1.

Comparison of floral morphological characteristics according to the geographical division of Daphne kiusiana. The western populations show more and larger flowers, whereas the eastern populations show fewer and smaller flowers. The longitudinal sections of the flowers illustrate that the levels of herkogamy are significantly different between the western and eastern populations. (A) inflorescences; (B) habitat; (C) longitudinal section of the flower.

Fig. 2.

Comparison of genetic diversity between the two lineages of Daphne kiusiana. (A) Boxplots showing mean allelic richness (AR), expected heterozygosity (HE) and haplotype diversity (Hd). The mean within-population values for each lineage are shown as red lines. Each value of the pooled lineages is marked with an asterisk. (B) Histogram showing the observed heterozygosity (H_O) at the level of individuals for each lineage. LI, lineage I (west); LII, lineage II (east).

Fig. 3.

Genetic composition of Daphne kiusiana populations. (A) STRUCTURE clustering using 16 microsatellite loci (K = 2) and barrier analysis based on Monmonier’s algorithm and significance, tested using means of 1000 bootstrap matrixes of D_A genetic distance (Nei et al., 1983). The rate of change in the log-likelihood probability and ΔK based on the estimated number of genetic clusters (K) is shown in the upper-right inset. (B) Geographical distribution of three chloroplast haplotypes based on 16 non-coding cpDNA regions. The grey shading represents exposed coastal areas and sea basins during times of glacially induced alterations in sea level during the Late Pleistocene.

Fig. 4.

Phylogenetic relationships (neighbour-joining trees) for nine populations of Daphne kiusiana based on (A) Nei’s chord distance (D_A) and (B) the proportion of shared alleles (D_ps). Values in the tree branches are percentage bootstrap values estimated from 1000 reiterations. Bootstrap support at internodes are shown if values are >50 %.

Fig. 5.

Inferred demographic history for Daphne kiusiana. (A) Trajectories of effective population size change in lineage I and lineage II predicted by the size reduction model. Lines and coloured areas indicate the posterior mode and 95 % highest posterior density. (B) The best supported population divergence model in ABC analysis of the genetic data sets. N_CUR, current effective population size; N_ANC, ancestral effective population size; T, event time for population size change; T_DIV, divergence time.

Fig. 6.

Potential distributions of Daphne kiusiana predicted by ecological niche modelling. The potential range during the LGM was averaged from three general circulation models. (A) Present; chloroplast haplotype network based on 16 non-coding cpDNA regions of var. kiusiana and var. atrocaulis. (B) LGM; ML tree is boxed in grey. HKC 1 indicates the main track of the Kuroshio Warm Current (proposed by Ujiie et al., 2003; Kao et al., 2006; Zheng et al., 2016). The scale bar in the ML tree represents the average number of nucleotide substitutions per site. HKC 2 is the hypothetical Kuroshio Warm Current (proposed by Vogt-Vincent and Mitarai, 2020).

Fig. 7.

Boxplots expressing morphological variations in two Daphne kiusiana lineages (LI and LII). The box signifies the distribution of the 25–75 % quartiles, and the median is represented by a horizontal line within the box. The ends of the vertical lines indicate minimum and maximum data values, respectively. Mean values are represented by a yellow line, with the value shown next to the line. The differences between the two lineages were significant for each respective trait where P < 0.001 (F test).

Fig. 8.

Scatter plots showing correlations between herkogamy and floral morphological traits of Daphne kiusiana. Blue and red circles represent the data of lineage I (LI) and lineage II (LII), respectively. Grey straight line, fitted values; shaded area, 95 % confidence interval; ***, P < 0.001.